The purpose of this study was to investigate the developmental change of celf-concept of male amd female students from grade 5, 8, 11, and sophomore in collleges. To achieve the purpose, the following specific objectives were set forth. a. Trait salience by grade levels in the free-answered positive and negative self-evaluations. b. Age differences in the means of the global and seperate aspects of self-concept. c. The impacts of moral judgment and causal attributions for success and faliure on academic self-concept and general self-worth. d. The combined and relative effects of selected variables such as SES, age, moral Judgment, causal attributdions, perceived recent achievement on academic self-concept and general self-worth. The instruments used for the study were open-ended questionnaire 'who am l'(WAI) designed to describe about his(her) own good and bad points that individual subject has thought and felt, structured questionnaire of self-concept(SCQ) with 85 items designed to measure seven sub-area of self-concepts, structured questionnaire of causal attributions for success and failure(CASF) with 40 items, and Korean DIT(KDIT)of short form with three dilemma stories. The instruments of WAI, SCQ, and CASF were administered to all the subjects and KDIT to the subjects of grade 8, 11, and sophmore. The major findings were as follows: a. The dimensions of free-answered self-evaluations contained general self, physical self, social self, family self, moral-ethical self, personal self, self of intellectual ability, self of specific skills, and others. b. The negative responses in free-answered self-evaluations were higher than the positive responses for 8th and above graders. c. As the grade(age)got higher, the responses of general and abstract aspects of self (e.g. general self and personal self) in free-answered self-evaluations were increased, while the responses of concrete and observable aspects of self(e. g. body images) were decreased. Both academic self-concept and general self-worth measured by SCQ showed lower but significant positive correlations with DIT P scores for higher graders (11th and sophomore). These results suggest that age differences of self-concepts are clolsely related to the development of cognitive structure. d. The means of global self-concept and most of sub-area self-concepts measured by SCQ were significanty decreased for 8th and 11th graders (secondary school students). It is suggested that the period of junior high school may be the turning point to change the self-concept more negatively. e. Both academic self-concept and general self-worth showed significant positive correlations with success/internal attribution, and significant negative correlations with success/external attribution for 5th, 11th graders and sophomore. f. By the results of MRA(multiple regression analysis), perceived recent acheivement appeared to be most powerful in academic self-concept for all the graders. It was found that for 5th and 11th graders, second important variable for academic self-concept was success/effort attribution and for sophomore, moral judgment(DIT P scores). For gernal self-worth, perceived achievement appeared to be most powerful for lower graders(5th and 8th)and success/effort attribution for higher graderst(11th and sophmore). It was found that the variable of second powerful effect on self-worth was success/effort attribution for 5th graders and moral judgment for sophomore. Neither SES nor age had additional effect on both academic self-concept and self-worth for total sample.